Machine Learning Times
Machine Learning Times

CONTINUE READING: Access the complete article in data-informed, where it was originally published.  

5 years ago
How to Harness Predictive Analytics and Become a Big Data Dynamo


Predictive analytics has come a long way and, in an era defined by the ever-increasing influx of data and heightened customer demands, businesses no longer can deny its strategic importance.

Industries such as insurance, financial services, and retail have used predictive analytics for decades, while others are just getting started. So what’s new? Predictive analytics now is being used to support day-to-day business operations and decision making rather than only special, retrospective projects.  Companies that use predictive analytics effectively can glean forward-looking insights that enable them to spot new business opportunities and innovate more quickly.

So why aren’t more companies doing it? Many simply don’t know where to start. Becoming a big data dynamo in your organization doesn’t require a complete rethink and change in how things are done. However, while business leaders agree on the importance of a data-driven approach to survive the next decade, an overwhelming number admit that, at the core, they still struggle with information overload and deriving actionable insights from data they already possess.

USE CODE PATIMES16 for 15% off Predictive Analytics World Conference pass.

Using predictive analytics, while difficult, is possible and also increasingly necessary to compete effectively today. In fact, in a recent study by Capgemini, 65 percent of respondents agreed that their business runs the risk of becoming irrelevant if they do not embrace big data.

Here are a few tips to position your company for predictive analytics success:

Focus on a Place to Start

With a sea of potentially useless data, narrow down your options and find the right area of your business to get started. CapGemini studies confirm that the number-one guiding principle to harnessing success with big data is to focus on solutions supporting your primary business objectives. Depending on your function, some considerations are market planning, account intelligence, and optimizing operations to streamline processes.

  • Market planning. To spot the right market opportunities, use simulations to blend economic data with your business-performance data to determine who is most likely to buy so you can focus efforts and build and deploy resources most effectively.
  • Account intelligence. Consumer brands born on the Internet, such as Amazon or Alibaba, excel at market basket analysis – analyzing customer purchasing behavior to figure out what they might buy next – but this is relatively new in the B2B space. Basically, you can determine which accounts have the greatest propensity to buy based on which purchases are likely to go together. Retailers use market basket analysis for promotions and targeted recommendations. Similarly, B2B companies can use internal and external tools for such analysis to amp up traditional lead-generation efforts.
  • Optimizing Operations. Ask yourself: Which of your day-to-day operational tasks can be done smarter? At EMC, for example, we noticed that there were more contract renewal opportunities than we had sales reps to make phone calls, so we use analytics to prioritize the highest-value renewals and direct the focus of sales reps.

Collect the Right Data

Some people believe those with the most data will win. My experience is that those with the right data win. The quality of data sources must be the top priority when launching an analytics function. Start small and look to incorporate both internal and external data sets.

Make Smart Hiring Choices to Build the Right Bench

It takes more than data scientists crunching data to be successful. Having the business perspective to develop actionable insights is essential. Build a diverse analytics team with a wide variety of skill sets, including data officers, analysts, engineers, scientists, and consultants.

Click to enlarge

As you build a “data bench,” there are three or four roles to consider. The first is a data analyst, who is intimately familiar with how to extract and transform data for its intended purpose. Second is the data engineer, the person who knows how the data is being captured, the servers where the data is located, and the tools that are required to extract data for analysis. Third are the data scientists, who can create a profile or do clustering analysis on data. Finally, segment experts or consultants can contextualize the findings and deliver recommendations that are compelling to senior leaders.

Manage Organizational Change

Incorporating analytics into the decision-making process is not always welcomed by all stakeholders, so it’s important to manage the change strategically. You can ensure a smoother transition through the following steps:

  • Encourage engagement. Pilot programs and focus groups can help involve stakeholders in the process of building out an analytics function. Generally, folks want to be part of the solution rather than being given the final output. Early participation is a huge lift in driving change.
  • Enlighten and inform. Even the most sophisticated machine-learning models are grounded on some foundational data principles. Exposing some or all parts of a predictive model delivers transparency, which leads to trust, which leads to adoption.

Overwhelmed? Get Inspired, Get Going, Get Organized

CONTINUE READING: Access the complete article in data-informed, where it was originally published.

Author Bio:

John Smits is Chief Data Officer for EMC Corporation’s Global Business Operations. John leads EMC’s Business Insights and Analytics team in the development of market segmentation, customer intelligence, and business performance analytics that are enabling its digital sales transformation with big data insights. Check out John’s Wise Practitioner – Predictive Analytics Interview Series.

Leave a Reply